Multidrug-resistant (MDR) Gram-negative 'superbugs' are rapidly spreading around the world, and polymyxin B and colistin (polymyxin E) are often the only effective antibiotics. Since polymyxin B was released in the 1950s, its pharmacokinetics, pharmacodynamics, toxicodynamics (PK/PD/TD) have never been defined. Recent pharmacological research on polymyxins has predominantly focused on colistin methanesulfonate (CMS, an inactive prodrug of colistin) and demonstrates that CMS has significant limitations. Thus, polymyxin B is increasingly being viewed as the preferred polymyxin. Unfortunately, recently developed scientifically-based dosing recommendations for CMS cannot and should not be applied to polymyxin B, as the latter is administered as its active entity. Therefore, it is essential to determine the PK/PD/TD of polymyxin B in critically-ill patients, refine optimal dosage regimens, and develop the user-friendly adaptive feedback control (AFC) clinical tool. The Specific Aims are: 1) To develop a population PK model for polymyxin B; 2) To investigate relationships between the PK of polymyxin B, duration of therapy and patient characteristics, with the development and timing of nephrotoxicity; and to use next-generation proteomics to identify the most predictive biomarker(s) of polymyxin B associated nephrotoxicity; and to develop the population PK/TD model; 3) To establish the relationships between polymyxin B PK, bacterial susceptibility and patient characteristics, with the probability of attaining and time to achieving clinical and bacteriological outcomes; and 4) To employ the models from Aims 1-3 and Monte Carlo simulation to develop scientifically-based dosage regimens of polymyxin B and to develop an AFC algorithm for future individual patients. Research Design: Patients being treated with intravenous polymyxin B will be identified at three clinical sites in the USA and one in Singapore. Patients (n = 250) will have blood collected at various times surrounding a dose of polymyxin B between days 1 and 5 of therapy. Development of nephrotoxicity, clinical response, and bacteriological response will be examined. Total and free plasma concentrations of polymyxin B will be determined. Bacterial isolates will be examined for the emergence of polymyxin resistance. The relationships between polymyxin B PK, PD and TD end-points (e.g. clinical and bacteriological responses, development of toxicity and resistance) will be assessed using pharmacometric analyses. Finally, the obtained information will be used to apply Monte Carlo simulation to examine the impact of various patient characteristics and other factors on polymyxin B PK, PD and TD, in order to establish optimal dosage regimens and AFC algorithms for individual critically-ill patients. Significance: No new antibiotics will be available for Gram-negative 'superbugs' for many years. This landmark multicenter study will provide essential information for optimizing polymyxin B use in critically-ill patients, while minimizing resistance and toxicity. This proposal aligns perfectly with the NIAID priority To teach old drugs new tricks and the recent Executive Order of the White House to combat antibiotic resistance.